HP Oracle Database Machine - IOUG

Download Report

Transcript HP Oracle Database Machine - IOUG

The Best of Oracle Open World 2008
Barry Hodges
Senior Solution Architect,
Sales Consulting, Oracle NZ
A Different Perspective
• Just being there …
• Opportunity to liaise with subject matter
experts and product managers
• Input into the road map
• Pre-arranged agenda adapted as new
opportunities presented themselves
Setting the scene …
Data Warehouses Growing Rapidly
Tripling In Size Every Two Years
Size of the Largest Data Warehouses
Terabytes of Data
1000
800
600
Actual
Projected
400
200
1998 2000 2002 2004 2006 2008 2010 2012
Source: Winter TopTen Survey, Winter Corporation, Waltham MA, 2008.
Data Warehouses Start Slowdown at 1TB
10 Hours
Table
Scan
Time
5 Hours
Typical
NAS
Typical
Mid-Range
Array
Typical
High-End
Array
1 Hour
Table Size
1TB
10 TB
100TB
The Performance Challenge
Storage Data Bandwidth Bottleneck
• Current warehouse deployments often have bottlenecks
limiting the movement of data from disks to servers
• Pipes between disks and servers are 10x to 100x too slow
for data size
Solutions To Data Bandwidth Bottleneck
• Ship less data through the pipes
• Add more pipes
• Make the pipes wider
Solution Approaches
• “Brainy Software” approach - use smart
database software to minimize the need for
hardware
– Smart Scan, OLAP, Bitmap Indexing, Join indexing,
Materialized Views, Result Caches, Range
Partitioning, etc.
• “Brawny Hardware” approach - use powerful
hardware to perform brute-force scans and joins
– Brawny hardware approach only works if the rate of
scanning data scales with size of data
What We Announced …
HP Exadata Storage Server Hardware
The Core Building Block
• Exadata Ships Less Data Through Pipes
– Query processing is moved into storage
to dramatically reduce data sent to
servers while offloading server CPUs
• Exadata has More Pipes
– Modular storage “cell” building blocks
organized into Massively Parallel Grid
– Bandwidth scales with capacity
• Exadata has Bigger Pipes
– InfiniBand interconnect transfers data 5x
faster than Fibre Channel
Exadata Moves a Lot
Less Data a Lot Faster
HP Exadata Cell Building Block Details
• Storage Server (Cell) is building block of Exadata Storage
Grid
• Cells ship with all hardware and software components preinstalled
• Runs Exadata Storage Software, Oracle Infiniband protocol,
Oracle Enterprise Linux 5.1, and HP hardware management
software
• Hewlett Packard is exclusive hardware provider
• Absolutely no custom hardware - all parts off the shelf highvolume
Cell Hardware Details – HP DL180G5
Redundant 110/220V
Power Supplies
2 Intel 2.66 Ghz Xeon
Quad-core Processors
P400 Smart Array Disk
Controller card
- 512M battery backed cache
12 x 3.5” Disk Drives
- 7200 RPM 1000GB SATA
- 15000 RPM 450GB SAS
8 GB DRAM
Lights Out 100c –
Management Card
Infiniband DDR
dual port card
SAS or SATA Disks in Exadata Servers
• Choice of either
– 300 GB 15,000 RPM Serial Attached SCSI (SAS) disks
– 1 TB 7,200 RPM Serial Advanced Technology Attachment (SATA) disks
• Choose SAS Based Servers for High Performance
SAS Advantages
SAS
SATA
Advantage
Throughput (MB/s)
1,000
750
1.33X
Average Seek Time (ms)
3.5
7.4
2.11X
Disk level read errors (per year)
6.3
63
10.00X
Years to disk failure
15.2
11.4
1.33X
• Choose SATA Based Servers for High Capacity
SATA Advantages
SAS
SATA
Advantage
Capacity (TB RAW)
3.6
12
3.33X
Capacity (TB Uncompressed User Data)*
1.0
3.3
3.33X
* Computed after mirroring, log files, indexes, undo and temp tablespaces
SAS or SATA Disks in Exadata Servers
• Choice of either
– 450 GB 15,000 RPM Serial Attached SCSI (SAS) disks
– 1 TB 7,200 RPM Serial Advanced Technology Attachment (SATA) disks
• Choose SAS Based Servers for High Performance
SAS Advantages
SAS
SATA
Advantage
Throughput (MB/s)
1,000
750
1.33X
Average Seek Time (ms)
3.5
7.4
2.11X
Disk level read errors (per year)
6.3
63
10.00X
Years to disk failure
15.2
11.4
1.33X
• Choose SATA Based Servers for High Capacity
SATA Advantages
SAS
SATA
Advantage
Capacity (TB RAW)
5.4
12
2.22X
Capacity (TB Uncompressed User Data)*
1.5
3.3
2.22X
* Computed after mirroring, log files, indexes, undo and temp tablespaces
Traditional Scan Processing
What Were
Yesterday’s
Sales?
Oracle
Database Grid
Select
sum(sales)
where Date=
’20-Oct’ …
SUM
Storage Array
Retrieve
Entire
Sales Table
Exadata Smart Scan Processing
What Were
Yesterday’s
Sales?
Oracle
Database Grid
Select
sum(sales)
where Date=
’20-Oct’ …
SUM
Exadata
Storage Grid
Retrieve Sales
Amounts from
Oct 20
Smart Scan Functionality
• Join filtering
– Star join filtering is performed within Exadata storage cells
– Dimension table predicates are transformed into filters that are
applied to scan of fact table
• Backups
– I/O for incremental backups is much more efficient since only
changed blocks are returned
• Create Tablespace (file creation)
– Formatting of tablespace extents eliminates the I/O associated with
the creation and writing of tablespace blocks
• Smart scans are transparent to the application
– No application or SQL changes required
– Returned data is fully consistent and transactional
InfiniBand Throughput
Single Connection Throughput
1400
1200
1000
MB/sec 800
3x slower
600
400
12x slower
200
0
Gigabit Ethernet
4Gb Fibre
20Gb Infiniband
• Graph shows throughput achieved in real-world deployments
– Infiniband is held back by PCIe 1.0 x8 bus on typical host systems
Massively Parallel Storage Grid
• Scalable
– Scales to hundreds of storage servers
– Data automatically distributed across storage
servers by ASM
– Transparently redistributed when storage
servers are added or removed
– Data bandwidth scales linearly with capacity
• Available
– Data is mirrored across storage servers
– Failure of disk or storage server transparently
tolerated
• Simple
– Works transparently - no application changes
16 GB/sec
8 GB/sec
…
4 GB/sec
Exadata bandwidth scales
linearly with capacity
HP Oracle Database Machine
Pre-Configured High Performance Data Warehouse
• 8 HP ProLiant DL360 Oracle Database
servers
– 2 quad-core Intel Xeon, 32GB RAM
– Oracle Enterprise Linux
• 14 Exadata Storage Cells (SAS or SATA)
– Up to 21 TB uncompressed user data (SAS) *
– Up to 46 TB uncompressed user data (SATA) *
• 4 InfiniBand switches, 1 Gigabit Ethernet
switch
• Pre-installed, pre-configured
Add more racks for unlimited scalability
* Computed after mirroring, log files, indexes, undo and temp tablespaces
Software Components
• Required Software
–
–
–
–
Oracle Database 11g Enterprise Edition
Oracle Exadata Storage Server Software
RAC
Partitioning Option
• Recommended Software
–
–
–
–
–
–
–
Advanced Compression Option
Enterprise Manager Packs: Diagnostics, Provisioning, Tuning
OLAP Option
Data Mining Option
Advanced Security Option
Real Application Testing
Oracle Business Intelligence Enterprise Edition Plus
Build your own DB Machine
Single Instance
RAC
InfiniBand Switch/Network
* Currently certifying x86 OEL running Oracle 11.1.0.7
Exadata Performance Scales
10 Hour
Table Scan Time
Typical
Warehouse
5 Hour
1 Hour
Exadata
1TB
10 TB
100TB
Table Size
M-Tel Exadata Speedup
10X to 72X Performance Improvement
28x
Average
Tablespace Creation
Speedup
Index Creation
Handset to Customer Mapping Report
CRM Customer Discount Report
CRM Service Order Report
Warehouse Inventory Report
CDR Full Table Scan
0
10
20
30
40
50
60
70
80
Giant Eagle Exadata Speedup
3X to 20X Performance Improvement
Merchandising Level 1 Detail:
Period Ago
Merchandising Level 1 Detail:
Current - 52 weeks
Supply Chain Vendor - Year - Item
Movement
Merchandising Level 1 Detail by
Week
Materialized Views Rebuild
Date to Date Movement
Comparison - 53 weeks
16x
Prompt04 Clone for ACL audit
Average
Sales and Customer Counts
Speedup
Gift Card Activations
Recall Query
-
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
45.0
50.0
Database Machine Comparison
HP Oracle
Database
Machine
Netezza
10100
Teradata
2550
Footprint
1 rack
1 rack
1 rack
User data
21 TB
12.5 TB
12.6 TB
108 x 400GB disks
144 x 300GB disks
4 DB Cores (?)
32 DB Cores
108 Storage Cores*
0 Storage Cores
112 Cores*
32 Cores
1Gb/sec Ethernet
1 Gb/sec BYNET
108 GB
128 GB
Proprietary
Proprietary**
Disks
Database cores
Storage cores
Total cores
Interconnect
Memory
HW Architecture

168 x 450GB disks
64 DB Cores

112 Storage Cores
176 Cores

20Gb/sec Infiniband

368 GB

Open

* Netezza 10100 uses PowerPC CPU’s (less powerful than Intel Xeon cores)
** Teradata BYNET Interconnect is proprietary
Data Protection Solutions
• All single points of failure eliminated by the architecture
• Hardware Assisted Resilient Data (HARD) built into storage
– Prevent data corruption before it happens
• Data Guard provides disaster protection & data corruption protection
– Automatically maintained second copy of database
• Flashback provides human error protection
– Snapshot-like capabilities to rewind database to before error
• Recovery Manager (RMAN) provides backup to disk
– Archiving and corruption protection
– Can be used with Oracle Secure Backup (OSB) or third party tape backup
software
• These work just as they do for traditional non-Exadata storage
– Users and database administrator use familiar tools
Exadata Co-Existence and Migration
• Databases can be concurrently
deployed on Exadata and
traditional storage
– Smart Scan capability only works if the
whole tablespace is located on
Exadata storage
• Online migration if currently using
ASM and ASM redundancy
• Migration can be done using
RMAN or Data Guard
Database
Server
Exadata
Non-Exadata
Online Migration
The HP Oracle Database Machine
How Much?
US$24,000 / HP Oracle Exadata Storage Server
US$120,000 / Oracle Exadata Storage Server
Software (US$10,000 / disk)
US$650,000 / HP Oracle Database Machine
US$1,680,000 for Oracle Exadata Storage
Server Software (US$120,000 / Storage Server)
* Database 11g Enterprise Edition, RAC and Partitioning Option are
required and must be licensed in addition to that above.
* North America Prices. NOT inclusive of shipping, local taxes, etc
The HP Oracle Database Machine
Sales / Support Model
– System Delivery
– Hardware Service
– System Sales
– System Support
Oracle Technology Sales Manager is the single point for sales
Oracle is single point for Support
The HP Oracle Database Machine
Where Can You Try It?
• NZ TBC
– Singapore already
– Sydney, Australia scheduled
– NZ to be confirmed
A queue has already formed!
The HP Oracle Database Machine
How Is Oracle Prioritising Interest?
• Oracle Technology Sales Manager coordinates with
Oracle data warehouse specialist Sales Consultants
• We are looking for:
– Well-known companies or organisations who will be
references
– Oracle knowledgeable staff
– Performance challenges likely due to unbalanced hardware
– Budget and clear business need
– Query profile will benefit from Oracle Exadata Storage
Server Software
– Interest in full HP Oracle Database Machine configuration
Exadata Benefits
• Extreme Performance
–10X and more speedup for data warehousing
• Database Aware Storage
– Smart Scans
• Massively Parallel Architecture
– Dynamically Scalable to hundreds of cells
– Linear Scaling of Data Bandwidth
– Transaction/Job level Quality of Service
• Mission Critical Availability and Protection
– Disaster recovery, backup, point-in-time recovery, data validation,
encryption
The HP Oracle Database Machine
The next step in DW Hardware Solutions
Custom
Reference
Configurations
• Complete Flexibility
• Any OS, any platform
• Easy fit into a
company’s IT
standards
• Documented bestpractice
configurations for
data warehousing
Optimized
Warehouse
HP Oracle
Database
Machine
• Scalable systems
pre-installed and preconfigured: ready to
run out-of-the-box
• Highest performance
• Pre-installed and preconfigured
• Sold by Oracle
Resources
• Oracle.com:
http://www.oracle.com/exadata
• Oracle Exadata Technology Portal on OTN:
http://www.oracle.com/technology/products/bi/db/exadata
• Oracle Exadata white papers:
http://www.oracle.com/technology/products/bi/db/exadata/pdf/exad
ata-technical-whitepaper.pdf
http://www.oracle.com/technology/products/bi/db/exadata/pdf/migr
ation-to-exadata-whitepaper.pdf